Introduction

Deficiencies in children’s growth and nutrition are multifactorial in origin, but potential contributors include social inequalities in food distribution [1] and gender inequality [2]. Intimate partner violence [IPV], defined as “behavior by an intimate partner or ex-partner that causes physical, sexual or psychological harm, including physical aggression, sexual coercion, psychological abuse and controlling behaviors” [3] may also influence children’s nutrition outcomes. Different types of IPV are physical IPV (which includes slapping, hitting, kicking and beating [4]), sexual IPV (which includes forced sexual intercourse and other forms of sexual coercion, whether using physical force or other forms of coercion) [4], emotional or psychological IPV (which includes insults, belittling, humiliation, and intimidation [4]), and controlling behaviors (which includes actions such as a partner limiting respondent’s contact with family or friends or monitoring and restricting money management, movement, education, employment or medical care) [5, 6]. Male perpetration of IPV against women and girls is highly prevalent globally [7], and in fragile settings and humanitarian emergencies, the risk factors increase [8]. A 2011 systematic review of population-based studies focused on prevalence of all forms of violence against women and girls in complex emergencies indicated that IPV is one of the most common forms of violence experienced by women and girls in humanitarian contexts [9]; limited evidence confirms that IPV prevalence increases in conflict-affected areas compared to non-conflict settings [10].

Children–defined as anyone from birth until age 18, as per international definitions–can be indirectly exposed to IPV, which can include witnessing, being aware of, or secondarily affected by the presence of violence against a maternal caregiver in the household. Children’s indirect exposure to IPV is associated with lower rates of vaccination and with later adolescent and adult risk-taking behaviors [11]. There is robust literature on indirect IPV exposure and children’s adverse mental health outcomes, such as externalizing symptoms, (e.g. aggression, impulsivity, or anti-social behavior) [11]. Evidence indicates that pre-school aged children who witness IPV in childhood may experience detrimental effects on their later mental, social and physical health during adolescence and adulthood. Some children develop traumatic stress symptoms, mood anxiety disorders, depression, alcohol use, and have higher rates of violence perpetration [12–15]. A recent review of IPV against a maternal caregiver and child development outcomes in 11 low- and middle-income countries [LMICs] identified associations between IPV exposure and literacy, numeracy, and cognitive, physical and socioemotional development for children aged 36 to 59 months [16]. For example, a study conducted in South Africa showed that emotional IPV was associated with lower language development, motor development and cognitive scores in children at age 2, while physical IPV was associated with lower motor scores [17], and in a study conducted in Kenya, maternal exposure to IPV was associated with poorer child behavioral outcomes [18]. Girls and boys may experience different patterns of direct and indirect exposure to IPV, however, the limited research exploring these differences is mixed [19, 20], and there are also gender differences in mental health impacts of childhood IPV exposure [21]. The evidence-base concerning the impacts of childhood indirect exposure to IPV on children’s health outcomes is outside of mental health disorders and risk-taking behaviors is limited.

Nutritional vulnerabilities pose particular challenges in countries defined as ‘fragile or extremely fragile,’ which can be defined as countries “where the state power is unable and/or unwilling to deliver core functions to the majority of its people: security, protection of property, basic public services and essential infrastructure” [22]. Countries where there is national and/or regional fragility are more likely to have high levels of anemia, and those countries deemed as extremely fragile are more likely to have high levels of anemia as well as high levels of stunting and wasting (both markers of chronic malnutrition) [23, 24]. For example, of the estimated 45.4 million children under 5 globally who suffer from wasting, one in four live in humanitarian contexts [24]. Existing evidence gaps illustrate that nutrition indicators are especially challenging to study and address in fragile settings. Indeed, Yount et al.’s review, published a decade ago, focused on presenting a possible framework for pathways between IPV and child growth and nutrition in the first 36 months, primarily included studies based on data gathered in HIC settings, and, to a much lesser degree, LMIC and “fragile” settings [25].

Several evidence syntheses have addressed the question of the association between children’s indirect exposure to IPV (against a maternal caregiver) and child nutrition outcomes. Yount et al.’s review found that IPV “may affect early childhood growth and nutrition through biological and behavioral pathways,” and that the strongest evidence available concerned the association between IPV and low birth weight [LBW], an association this review indicated is likely mediated by maternal prenatal risk behaviors, mental health and poor weight gain during pregnancy [25]. Other relevant reviews have focused on in-utero exposure to IPV rather maternal experience of IPV at other time points; Hill et al. found associations between IPV during pregnancy and LBW and preterm birth [26] and Donovan et al. found associations between IPV during pregnancy and pre-term birth, LBW and small for gestational age babies [27]. Experience of IPV can influence maternal care practices relating to child nutrition; a review of observational studies found that 8 of 12 included studies reported an association between IPV and breastfeeding practices (lower breastfeeding intention, breastfeeding initiation and exclusive breastfeeding) [28]. These existing evidence analyses focus on specific periods of experiencing IPV (in-utero), or a limited range of nutrition indicators (LBW, breastfeeding). To address existing evidence gaps, we conducted a rapid evidence assessment [REA] focusing on all forms of indirect experience of IPV, during any exposure period, and including nutritional/growth outcomes reported as a) fetal growth (measured indirectly with birth weight), b) infant feeding practices, as well as indicators of c) child growth (using anthropometry) and d) nutrient blood markers in order to examine acute and chronic malnutrition during infancy, early childhood (0 to 39 months, including those focused on children under 2 or under 5 years), and middle childhood (after the early growth spurt). Examining pathways concerning children’s indirect exposure to IPV against a maternal caregiver and childhood nutrition and growth can highlight how IPV adversely impacts child health and inform policies and programs at the nexus of IPV and nutrition. Our objective was to review and synthesize the current quantitative literature describing the associations between IPV against a maternal caregiver and children’s nutrition outcomes, with a particular focus on those studies carried out in LMICs and fragile settings.

Methods

The REA was conducted utilizing an adapted systematic review methodology [29]. Specifically, title/ abstract review was conducted by a single reviewer, and inclusion/ exclusion criteria (specifically, regarding the location of studies for inclusion) was determined iteratively. We developed and implemented a structured search of three databases: Medline, Embase, and Global Health. The initial search was conducted in July 2021, and updated in October 2022. The full search strategy is included in S1 Appendix. For further details on search methodology, see [blinded for peer review] et al. In brief, the search terms included the following fields: i) intimate partner violence, ii) nutritional outcomes, and iii) quantitative study design, with specific terms and MeSH headings tailored to each database.

The eligibility criteria were structured using the PECO framework–Population, Exposure, Context and Outcome [30] (see Table 1). The timepoints considered were during infancy, early childhood (including those outcomes specific to children under 2 or under 5 years), middle childhood and adolescence (after puberty or above age 10 years). Other inclusion criteria were that the manuscript was published in peer-reviewed literature, published in English or Spanish, after the year 2000. Spanish was chosen as a second language for the review due to the language capacity of the team. Any quantitative methodology studies were included, along with mixed methods studies if quantitative findings were separately reported.

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Inclusion criteria.

https://doi.org/10.1371/journal.pone.0298364.t001

All records identified through the database searches were downloaded to Covidence, a systematic review software. Screening occurred in two stages: (i) title and abstract; and (ii) full text review. Title/ abstracts were each screened by one reviewer [one of SBC, MOG, LV or SRM]. During the full text screening, sources were assessed independently by two reviewers vis-à-vis the inclusion/ exclusion criteria [two of SBC, MOG, LV or SRM]. A third independent reviewer resolved any discord between the first two reviewers.

Results

There were 6493 sources imported to Covidence. Duplicates (1516) were removed, yielding 4977 citations for screening. During title and abstract screening, 4461 sources were excluded and 468 full texts were assessed for eligibility. During full text review, 450 sources were excluded for various reasons outlined in Fig 1. Specific characteristics of the 86 articles included are listed in Table 2. There were 57 cross-sectional studies, 23 prospective cohort studies, 5 case-control studies, one observational cohort study, and one prospective case-control study. About 50% of the included studies were published in the past five years, with 20 studies published in 2021 or 2022, indicating a rapidly expanding evidence-base. Study results are in displayed in S1 Table.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. PRISMA flow chart.

https://doi.org/10.1371/journal.pone.0298364.g001

Download:

• PNG
  larger image
• TIFF
  original image

Table 2. Study characteristics–Indirect pathway.

https://doi.org/10.1371/journal.pone.0298364.t002

Table 3 displays results from the quality assessment. All individual quality assessment results are included in S2 Table. The sampling approach was highly variable, with 39.5% utilizing probability-based sampling methods and just over half (53.5%) including convenience sampling of women (typically women attending prenatal care or women who had given birth in facility/ hospital). Studies that utilized probability-based sampling methods were secondary analyses of Demographic and Health Surveys [DHS] data. In 40.7% of studies had a response rate of more than 90%. In 50.0% (n = 43) of studies, there was no description of the response rate and/ or any differences in characteristics between respondents and non-respondents. Despite these limitations, a high proportion of studies were categorized by the review team as having utilized rigorous measurement of IPV, i.e. questions that address acts-based items, rather than general questions about IPV experience (91.9%), and a similarly high proportion (87.2%) of studies included a rigorous measure of the nutrition outcomes (e.g. Z scores).

Download:

• PNG
  larger image
• TIFF
  original image

Table 3. Quality assessment.

https://doi.org/10.1371/journal.pone.0298364.t003

The REA findings are reported here by child developmental stages in the following sections: Low Birth Weight, Feeding Practices including Breastfeeding, and Indicators of Undernutrition (i.e. stunting, or wasting/underweight). The associations between maternal experience of IPV (lifetime, past year, and during pregnancy, wherein the affected child was in utero) and child nutrition and growth outcomes are shown the harvest plot in Fig 2. IPV was assessed separately (i.e. physical IPV only, sexual IPV only, emotional/ psychological/ verbal IPV only and controlling behaviors only) and as combined IPV, defined as physical, sexual and emotional, physical and sexual, or combined IPV–physical, sexual and emotional, physical and sexual, or any combination of IPV types. For each type of outcome, we include details from select studies that illustrate the associations found.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. Harvest plot of maternal experience of IPV and child nutrition and growth outcomes.

Harvest plot depiction of associations between maternal exposure to IPV and child outcomes related to growth and nutritional status, by timepoint of exposure. The figure below the plot shows a life-course and outcome timeline. (a) Shows associations with low birthweight. (b) Shows associations with feeding practices including breastfeeding. (c) Shows associations with stunting, linear growth and growth velocity. (d) Shows associations with underweight/ wasting/ severe malnutrition. (e) Shows associations with lipid biomarkers. The blue bar in the graphs depict maternal lifetime exposure to IPV, the orange bar depicts exposure to IPV in the past year and the gray bar depicts exposure to IPV during pregnancy.

https://doi.org/10.1371/journal.pone.0298364.g002

Discussion

This REA represents an important update to prior reviews focusing on the association between a maternal caregivers’ experience of any type of IPV and children’s nutrition outcomes. While the global evidence-base remains biased towards evidence gathered in high-income contexts, our review indicates that the past decades of research and analysis has generated significant insights into the associations between IPV and various nutrition outcomes in LMICs in all regions.

In examining potential mechanisms for how IPV can contribute to each of the groups of outcomes, LBW, feeding practices and various indicators of child growth and malnutrition, there is scope for hypothesizing causal pathways between experience of IPV by a maternal caregiver and nutritional/growth outcomes. Based on the data gathered from multiple studies for this review, there is a suggestion of a linear impact of experience of IPV during pregnancy on physiological pathways affecting fetal growth. This is supported by the strong and consistent associations between mothers’ experience of physical or combined IPV and LBW, while the impact of emotional/psychological IPV or sexual IPV may work through an intermediary in the pathway to LBW.

Although thirteen studies assessed the association between maternal experience of any type of IPV and LBW in the past year, only one study found a positive association with physical IPV [46]. In contrast, positive associations were found in six of seven studies examining lifetime combined IPV and LBW [39–45]. Similarly, 19 of 25 studies found a positive association between maternal experience of any type of IPV during pregnancy and LBW. The notable difference between experience of any type of IPV during a lifetime and during pregnancy vs experience of any type of IPV in the past year indicate the importance of taking into account timing of when a maternal caregiver experiences IPV to understand causal pathways to child nutrition outcomes.

Our findings find a significant association between mother’s experience of IPV and breastfeeding. The evidence-base confirms the hypothesis that women who have experienced IPV face psychological or physical barriers to breastfeeding, known as the deficit hypothesis, as opposed to the compensatory hypothesis, whereby women who have experienced IPV may be more sensitive or responsive to their children’s needs and therefore have higher odds of early initiation of breastfeeding, exclusive breastfeeding and/ or continued breastfeeding than women not exposed to IPV [75]. Unfortunately, most studies did not specifically focus on IPV occurring between birth and the interview. This is an important period to examine in new research, given that the peak risk periods during which experience of IPV experience can impact adherence to recommended practices in breastfeeding could be immediately following delivery as well as during pregnancy. Breastfeeding is critical for child health and growth, and contributes improved long-term health, including cognitive capacity and survival [98–100]. Among its multiple benefits are protection from infections during infancy, as well as providing hygienic food in settings with limited potable water (especially relevant in humanitarian settings), in fragile settings, the life-saving protection of breastfeeding is thus especially critical. Conflicts, natural disasters and epidemics contribute to forced migration resulting in heightened food insecurity, limited access to clean water, and disruptions to basic services, leaving women and children especially vulnerable. In these settings, breastfeeding guarantees a safe, nutritious and accessible food source for infants and a protective shield against infectious disease and death [100]. Preventing, managing or mitigating IPV can improve breastfeeding practices. Initiation of breastfeeding promotion interventions that reach mothers even before conception and recognizing the impact of a maternal experience of IPV on her breastfeeding practices, can impact child health indicators throughout childhood. Although some literature shows that breastfeeding is affected by sex feeding preference, where the mother may breastfeed a male more than a female child [101, 102], no articles on feeding preference and IPV were identified that met the inclusion criteria for this review. In addition, we found a surprising lack of sex disaggregation of child nutrition outcomes, which prevents proper analysis on sex based preferential breastfeeding and other feeding practices.

Findings indicate that experience of IPV by a maternal caregiver is associated with impaired child growth and indicators of acute and chronic malnutrition. Despite some lack of comparability of associations, the studies included in this review suggest that maternal experience of IPV affects several growth indicators in preschool children, however, the literature is extremely limited in its examination of time intervals that can fully inform our understanding of any potential underlying mechanism. Despite this, the literature that includes stunting as an outcome does suggest that maternal caregivers’ lifetime experience of IPV is more predictive of stunting than past year experience of IPV. This is suggestive of a pathway consistent with evidence where lifetime IPV can lead to inadequate nutrition and reduced care practice. The literature reviewed suggested that maternal caregivers’ experience of physical alone, sexual alone or combined IPV posed a greater risk for a child’s stunting, rather than other types of IPV. For more acute growth indicators, other types of IPV (physical) and a shorter period (past year) appear to pose greater risk. To better examine the pathways through which stunting and underweight may result from maternal caregivers’ experience of IPV, further work should examine underlying potential mechanisms, including examination of feeding patterns. The finding of stunting and lifetime maternal experience of IPV, may be particularly amenable to earlier intervention and strategies for prevention. In particular, capturing reporting of lifetime experience shows an underlying chronicity of exposure and long developmental window for stunting, which may have a physiologically different impact.

Though few studies examined maternal caregivers’ experience of IPV alongside children’s nutrient biomarkers in blood, the results from the one study suggests a potential causal pathway for lipid dysregulation (at age 10) leading to metabolic syndrome and contributing to later obesity resulting from maternal caregivers’ experience of physical and sexual IPV [97]. One other study found an increase in cortisol levels for newborns where the maternal caregiver had experienced several types of IPV during pregnancy [65] and in another study, a higher cortisol level in the mother, as well as the baby, was associated with lower birth weight [103]. Notably absent from the literature were studies examining the impact of maternal caregiver IPV on a child’s anemia during early childhood when children are reliant on caregiver feeding. Similarly, examination of maternal caregiver exposure to IPV and child measures of lipid dysregulation during puberty and adolescence may further elucidate potential pathways contributing to a child’s less healthy food choices later in life.

Results of the REA indicate the importance of context in associations between maternal caregiver experiencing IPV and child nutrition outcomes. In some multi-country studies, associations were found in one context and not another. For example, Misch et al., reported significant associations between physical IPV and delayed initiation of breastfeeding in Tanzania, Malawi and Zimbabwe and no association in Ghana, Liberia, Nigeria, Zambia and Kenya. Further supporting this example they found an association between sexual IPV experience and delayed initiation in Zambia but not in other countries [71]. Despite consistent use of methodology, IPV recall period and measurement of both IPV and nutrition outcomes in this multi-country study, the variation in results between contexts indicates that important, and likely unmeasured, contextual factors influence the association between indirect IPV exposure and child nutrition outcomes. Variance in associations occurred in a number of multi-country studies. In addition, intra-country variation in single-country studies may be substantially masked by country-level analyses of DHS data, in particular.

The REA also identified significant but addressable knowledge gaps, one, in the examination of risk during time periods preceding occurrence of the outcomes, and secondly in considering time periods that would be physiologically and developmentally meaningful for the progression of the underlying processes measured in the outcomes chosen. For example, although a large number of studies examined risk associated with experience of IPV during pregnancy, no studies considered the relevance of timing during pregnancy considering fetal development when examining the association between experience of IPV against a maternal caregiver and a child’s LBW. Additionally, studies examining the impact of physical IPV did not differentiate physical IPV involving direct abdominal trauma. Similarly, in those studies examining experience of IPV and breastfeeding, there was no documentation of IPV occurring during the child’s infancy (i.e. post-natal). Studies examining stunting or growth velocity similarly did not disaggregate maternal caregivers’ experience of IPV after the child was born. Many of these limitations may be due to the reliance on secondary data from large standardized regional/ national surveys. Studies reporting primary data collection though smaller were generally more able to provide insight into underlying processes.

Other gaps in the evidence-base identified in the REA include that only one study considered impact on young children by considering feeding practices in children above 6 months old, and using the UNICEF indicators for minimal acceptable diet (a composite of minimal dietary diversity, and age specific minimal meal frequency) [82]. These data are available in DHS datasets, and there is potential to conduct several secondary analyses of DHS data, including cross-country comparisons, to illustrate the associations between IPV and minimal acceptable diet. Studies in infants, children, and adolescents should include measures of diet diversity and diet adequacy as this is understudied (or completely unstudied). Given our findings regarding associations of IPV with breastfeeding, as well as the evidence from the two studies that examined minimal acceptable diet, it is highly plausible that feeding/ patterns of dietary intake in children older than 6 months will also be affected by either direct or indirect exposure to IPV. Strategies to prevent or remediate poor growth in infancy could be directly informed by including measures of the impact of IPV on dietary adequacy in infants and children.

A notable limitation of the evidence-base is the reliance on cross-sectional study designs as they do not allow for follow up to other outcomes as the child grows. The measurement and categorization of confounders, and inclusion of confounders, varied widely between studies, limiting direct comparison of results even if studies were using the same exposure and outcome measures. While measurement of IPV across the evidence-base was generally of high quality, inconsistency in grouping of types of IPV and different use of recall periods (lifetime, past year and during pregnancy) across studies led to lack of comparability of many associations identified in the studies. The secondary data analyses have inherent limitations because the studies are not designed to examine the association of specific types of IPV against a maternal caregiver with specific child outcomes. Many of the datasets could be used to examine these potential pathways with more precision by restricting the analyses to those in whom the time period of measured exposure corresponds to, a time period that is biologically and plausibly relevant for the outcome of interest.

There are various programmatic implications resulting from the REA. The evidence highlights the potential for greater synergies across GBV, nutrition, and maternal and child health programs. By identifying entry points within nutrition programs to provide support to survivors and refer them to specialized GBV services, such programs could improve well-being for both women and children. Equally, there are also opportunities to integrate more information on nutrition into GBV services to ensure that women who experience GBV are able to access nutrition services when their children are at risk. The evidence provides further support for advancing GBV prevention, given the potential lifelong impacts of IPV on child nutrition, and therefore, health and wellbeing. Effective approaches, such as UNICEF’s Communities Care model, hold promise to reduce GBV and thereby also contribute to better nutrition outcomes [104]. The findings of the assessment illustrate that incorporating GBV considerations into nutrition policies and programming is relevant to the protection and promotion of child health. Taking action to proactively identify and mitigate GBV-related risks helps make nutrition programming more effective especially in humanitarian contexts where the risk of GBV is higher.

Limitations

The results of the REA should be interpreted in light of some limitations, including that we conducted single screening of titles and abstracts (a single reviewer determined whether to include or exclude), and that data extraction was conducted by a single reviewer for each article. This methodology is in line with recent guidance regarding how to adapt systematic review methodology to the purposes of a rapid evidence assessment [105]. A second team member checked the data extraction for completeness and accuracy, and title/ abstract screening was conducted after 10% of the sample was double-screened, to ensure inter-rater reliability and consistency in application of inclusion and exclusion criteria. Studies in languages other than English and Spanish were excluded from the review due to language limitations of the team. The most significant limitation in interpretation of the results is that we did not weight studies according to quality or sample size. This level of quality assessment was beyond the scope of the REA in terms of timeline and resources. The findings regarding type and direction of associations, and patterns of associations between specific types of IPV and specific nutrition outcomes, should therefore be interpreted with this limitation in mind. Despite these limitations, this review represents an important update to prior evidence synthesis of the associations between maternal IPV and child growth and nutrition. By capturing the past few years of research, which has seen an enormous growth in published analyses, represents a comprehensive and up-to-date synthesis of this literature.

Conclusion

Our review identified maternal caregivers’ experience of IPV to be directly associated with low birthweight and breastfeeding practices, and also significantly associated with impaired fetal, infant, and child growth and indicators of acute and chronic malnutrition. There are opportunities that could be more strategically leveraged to better understand how experience of IPV can impact child growth indicators. Future studies could be designed to specifically evaluate experience of IPV during periods of risk relevant to specific growth outcomes, and to disaggregate analysis of outcomes by child sex. Despite some lack of comparability of associations identified in the studies, and differing patterns between specific types of IPV and different nutrition outcomes, there is ample evidence that IPV negatively affects a broad range of child nutrition outcomes. Acknowledging IPV as a risk factor within nutrition and child health programs holds the potential to significantly improve the nutrition, health and well-being of both children and women. Linking maternal caregivers’ experience of IPV to child malnutrition and growth pathways creates a space for conversations around integrated program models and the specific need for more development, testing and refinement of these models in LMIC settings. Approaches including targeted IPV screening of women whose children are undergoing treatment for malnutrition through supplementary feeding programs, training pre-natal health care workers and lactation specialists to identify signs of IPV amongst women, and strengthening referral networks between health and nutrition services and IPV response services are all key actions that can be explored in LMICs.

Supporting information

Acknowledgments

We thank Lily Coll for her support in data extraction.